Sign In

Communications of the ACM


Privacy and Google Glass

View as: Print Mobile App Share:
Carnegie Mellon Associate Professor Jason Hong

Google Glass, the wearable headset computer that is just starting to trickle out, has seen a lot of press, much of it negative on the issue of privacy (see for examples The New York Times, LA Times, Guardian, and ABC News). As someone who has studied mobile computing and privacy for about a decade, I thought it would be useful to examine why there has been such negative sentiment, looking at people's concerns from two different perspectives, and see what lessons can be drawn. (For disclosure purposes, Google has funded my research several times, and will be sending my colleagues and I several units of Glass for research purposes).

Lessons from the Ubiquitous Computing Project

Let's start out by rewinding back to 1991, when Ubiquitous Computing was first introduced to the world at large. In the seminal paper The Computer for the 21st Century, Mark Weiser presented a grand vision where one day, computation, communication, and sensing would be enmeshed in the everyday world, and could be used to seamlessly support us in our daily lives.

Weiser led his team in developing new form factors for computation at three physical scales: tabs, pads, and boards. They also developed early forms of location-based services and context-aware computing, based on wearable badges that could be used to pinpoint one's location inside of a building. While this vision may seem like old hat now, one has to keep in mind that this work was started only a few years after the first Macintosh came out, and mobile phones were still bulky affairs that were yet to be popular.

The researchers at PARC were, by and large, very enthusiastic about the potential of ubiquitous computing. However, the popular press was decidedly not. There were a number of negative news articles, with titles like "The Boss That Never Blinks," "Orwellian Dream Come True: A Badge That Pinpoints You," and "You're Not Paranoid: They Really Are Watching You."

I suspect there were two reasons why there was a backlash. First, the PARC researchers hadn't built in any kinds of privacy protections into their system. While they knew that privacy was a concern from the start, they didn't have good ideas for how to address the problems. Furthermore, they had to devote a lot of effort on just making the technologies even work in the first place. However, when journalists asked the inevitable questions about privacy, the researchers didn't have any good answers. To some extent, this scenario is playing out in the exact same way for Google Glass. Many people are asking reasonable questions about privacy, and I haven't heard any solid responses yet about how those concerns will be addressed.

Second, there was an unclear value proposition for end-users. The PARC team was comprised mostly of people from technical backgrounds, and when talking with journalists, the discussion usually centered on how the technologies worked rather than what benefit they could offer people. Interestingly, the narrative seemed to shift when the researchers started framing things in the form of "Invisible Computing," talking about how these systems could support people as they go about their everyday lives. Google is doing better on this front, by pushing the fashion aspects of Glass, by having some concept videos of what Glass can offer, and by having lots of non-developers try out the system. However, Google still has a long way to go in conveying what value it can actually offer today to everyday people.

This notion of the value proposition has been seen in the success and failure of many groupware systems as well. Jonathan Grudin, a scientist at Microsoft Research, long ago observed that those who do the work in using a groupware system have to be the same as those who get the benefits, otherwise the system is likely to fail or be subverted. My privacy corollary is that when those who bear the privacy risks do not benefit in proportion to the perceived risks, the technology is likely to fail. Right now, the implicit narrative in the popular press is that many people could be surreptitiously monitored by users of Google Glass at any time, and do not perceive any kind of value in return. Unless this perception is changed, it is likely that there will continue to be negative perceptions of Glass outside of the core of early adopters.

How Expectations of Privacy Change Over Time

The second perspective I'll use for thinking about Google Glass and privacy is that of expectations.

One of my favorite papers about expectations is by Leysia Palen, an HCI researcher at University of Colorado, Boulder. In 2000, Palen presented a paper at the Computer Supported Cooperative Work conference looking at the behaviors and practices of new mobile phone users (PDF). One finding was that these new users were not particularly good at predicting what their own attitudes and behaviors would be a month after getting their first mobile phone. For example, before they got their mobile phone, many participants reported being annoyed at people who used their mobile phones while driving or for casual chat in public places like restaurants and movies. However, just a few weeks later, many of the participants were exhibiting those same behaviors. Another interesting finding was that participants who had more exposure to mobile phones through friends or colleagues were better at predicting how they would be using phones.

Many other technologies have faced similar changes in expectations over time. Warren and Brandeis' famous definition of privacy as "the right to be let alone" came about in part because new cameras in the late 19th century made it possible to take photographs in just several seconds, invading "the sacred precincts of private and domestic life." Kodak cameras fared no better on the privacy front in its early days:

The appearance of Eastman's cameras was so sudden and so pervasive that the reaction in some quarters was fear. A figure called the "camera fiend" began to appear at beach resorts, prowling the premises until he could catch female bathers unawares. One resort felt the trend so heavily that it posted a notice: "PEOPLE ARE FORBIDDEN TO USE THEIR KODAKS ON THE BEACH." Other locations were no safer. For a time, Kodak cameras were banned from the Washington Monument. The "Hartford Courant" sounded the alarm as well, declaring that "the sedate citizen can't indulge in any hilariousness without the risk of being caught in the act and having his photograph passed around among his Sunday School children."

Similarly, in the book America Calling, sociologist Claude Fischer documented the social history of the telephone. In one of my favorite passages, Fischer observed that at first, many people actually objected to having landline phones in their homes, because it

"permitted intrusion... by solicitors, purveyors of inferior music, eavesdropping operators, and even wire-transmitted germs."

While these examples may seem quaint by modern standards, they represented real concerns that people had at the time. In fact, it's worth pointing out that many of these same problems still exist (inferior music now comes in the form of the Muzak you listen to when put on hold). The main difference is that our expectations of how these technologies will be used have changed over many decades, as we have adapted via changes in our social norms and laws.

So there are two points here relevant for Google Glass. The first is that we all lack experience with how we might use wearable computers, and so it's very likely that most of our expectations will likely be off mark. The second is that expectations can change over time, as we learn to adapt to the technology and its affordances, but only if we start to see real value in it.

It's also worth pointing out that expectations can also change quite rapidly and dramatically. Perhaps the best example of this is with the introduction of Facebook's News Feed in 2006. Before News Feed was rolled out, you could only see a person's status updates by going to their individual profile pages. What News Feed did was aggregate all of those updates in a single place. When News Feed was first made public, people's initial reactions were predominantly negative, and often viscerally so. Many Facebook groups were formed denouncing News Feed, and Facebook CEO Mark Zuckerberg even went so far as to publicly respond to all of the negative press. Facebook stuck to its guns and continued to push News Feed. In the course of just a few months, a lot of the criticism died out as people saw value in News Feed and became used to it. Now, several years later, I seriously doubt you could find someone today who would want to give up News Feed.

Now, this doesn't mean that people's privacy concerns always change in the way you want, or the way you expect. There are also plenty of examples where products were killed or features rolled back due to serious privacy concerns. My main points here are that we all have little experience with wearable computers, expectations of privacy can change, and perceived value is a major factor in driving that change. However, there's still a very big gap in the community's understanding of the best ways of mitigating these kinds of privacy issues up front, and the best ways of managing and designing for those changes. All I can say for sure is to buckle up your safety belts, because Google Glass is just one of many of these kinds of big changes in computing we will likely see in the future, and it will be a wild, scary, crazy, and exciting ride.


Geoff Revill

Don't you worry that the change in control of our online self is completely in the hands of companies who are primarily answerable to shareholders in the pursuit of profit?

As being online becomes necessary to engage with society at large don't you think that the real worry for the future is not governments as per the NSA and the Snowden debacle...but instead companies holding vast detailed knowledge about individuals thoughts habits desires ... I realise Google and facebook puts the USA in the driving seat of societal understanding globally.....but as the dataset refines rapidly towards the point where anonymity is impossible in the dataset in there is no such thing as a secure system...we are heading for a major problem for our next generation?

Displaying 1 comment

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account